In determining connection relationships between archaeological fragments, archaeologists often consider the likelihood that they fail in identifying some of these relationships.
Given an initial fragmentation graph, this function aims to simulate the effects of such different likelihood values.
For each value in the likelihood
parameter, a new graph is generated by randomly removing the given proportion of edges from the input graph.
When generating series of comparable graphs with different likelihoods, the function internally resets the seed for random number generation, to ensure that edges are removed in the same order.
An error message is displayed if at least one of the likelihood
values is < 0 or > 1.